You are on page 1of 8

Available online at www.sciencedirect.

com

ScienceDirect
Transportation Research Procedia 3 (2014) 452 – 459

17th Meeting of the EURO Working Group on Transportation, EWGT2014, 2-4 July 2014,
Sevilla, Spain

Electric Vehicle Routing Problem with industry constraints: trends


and insights for future research
Anagnostopoulou Afroditia, Maria Boilea*, Sotirios Theofanisb, Eleftherios
Sdoukopoulosa, Dimitrios Margaritisa
a
Centre for Research and Technology Hellas (CERTH) / Hellenic Institute of Transport, Egialias 52, 15125, Marousi, Athens, Greece
b
Rutgers University, 100 Brett Road, Piscataway, NJ 08854, United States

Abstract

This paper presents and analyzes the one-to-many vehicle routing and scheduling problem with electric vehicles. Initially, focus
is given on the problem formulation and the restrictions imposed in practice are examined. EVRP is NP-hard in the strong sense
since it is natural extension of the well-known Capacitated Vehicle Routing Problem and requires substantial computational
effort for determining optimal or near optimal solutions for medium and large scale problem instances. A comprehensive
mathematical formulation is developed in order to model the EVRP and the multiple constraints appeared due to capacity
limitations, time window restrictions and the predefined charging level of the vehicles. In addition, recent trends for the EVRP
are analyzed producing valuable insights for future research regarding extra operational constraints, real-life data sets and
solution frameworks that embody approximation algorithms for an efficient and effective search of the solution space.
© 2014The
© 2014 TheAuthors.
Authors. Published
Published by Elsevier
by Elsevier B. V.is an open access article under the CC BY-NC-ND license
B.V. This
Selection and peer-review under responsibility of the Scientific Committee of EWGT2014.
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer-review under responsibility of the Scientific Committee of EWGT2014
Keywords: Electric vehicles; routing and scheduling; mathematical formulation

* Corresponding author. Tel.: +30-211-1069-591; fax: +30-210-6533-031.


E-mail address: boile@certh.gr

2352-1465 © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer-review under responsibility of the Scientific Committee of EWGT2014
doi:10.1016/j.trpro.2014.10.026
Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459 453

1. Introduction

In today’s economy, markets tend to become increasingly open and competitive. The existence of an efficient,
sustainable and environmental friendly freight transportation system is a vital factor for success. Environmental
concerns and management issues related to the flows of freight have become crucial concerns for modern companies
seeking to reduce energy consumption and the related production of greenhouse gas (GHG) emissions. The transport
sector is responsible for 30% of the CO2 emissions in EU while this share reaches the 40% in urban areas. For that
reason, the efforts for technological improvements and innovation in transportation have been increased during the
last decades and the electric delivery vehicles is the new trend in United States for many companies such as Fed Ex
and UPS.
There is no doubt that urban cargo distribution with electric vehicles is beneficial to the environment. Cargo EV’s
are emission-free (except brake and tire wear) and perceived as more silent in operation. CEP and pharmaceutics
services deliver with vans and small size trucks in regionally limited areas. Their daily distance range is below
140km. Therefore, EV’s might potentially replace Commercial Vehicles (CVs) in delivery services in the near future
(University of Duisburg-Essen, 2011). The cost competitiveness of electric trucks to conventional trucks is in the
long term one of the most critical aspects to achieve a large-scale implementation of electric transport. Davis and
Figliozzi (2013) did a study in the U.S. with three types of recent technology electric delivery trucks in order to
examine their competitiveness to conventional trucks under varying scenarios. The study showed that electric trucks
can be competitive in case the cost savings from the reduced operational cost are enough to overcome the
significantly high purchase costs. Other parameters of cost increase, according to Jüchter (2011), are the significant
costs for charging infrastructure in companies.
The process of recharging electric vehicles is an important issue as it greatly determines their use, both in terms of
availability and flexibility. Regarding freight transport the charging techniques that get most attention so far are slow
charging, fast charging, battery swop stations and inductive charging. Research in this field is so far addressing to
applications for passenger cars. Still an unsolved problem related to charging process is the lack of compatibility of
different chargers plugs. Standardization of charging plugs on a worldwide level is further discussed.
Limited literature discussing charging infrastructure for UFT indicates that since delivery vehicles tend to follow
the same route each day and return to its company premise every night the possibility to recharge slowly during the
night is a satisfying solution (Feng and Figliozzi, 2013; Ehrler and Hebes, 2012). The advantage of night charging is
that public charging infrastructure has a lower demand than through the daytime (Schönewolf , 2011) and in some
countries night current supply has a reduced cost. However, it occurred to be that the installation and operation of in-
house charging infrastructure was technically challenging, therefore higher costs than initially planned (Schönewolf,
2011). In case charging is to take place outside the transport company premises (public charging infrastructure)
logistic processes and vehicle routing need to be rescheduled (Schönewolf, 2011). Furthermore regaining energy
through public charging stations, batteries should be quickly charged without damage (Schönewolf, 2011).
The key factors that determine the competitiveness of electric goods vehicles are the purchase price, fuel price,
battery costs and lifetime and the vehicle utilization. Certainly rising conventional fuel costs in addition to falling
battery costs will boost EV’s market. At present electric vehicles are not competitive if routing constraints lead to the
purchase of additional vehicles above the required number of conventional vehicles. Therefore, current priority until
any future technological breakthrough is the optimization of delivery cargo EVs in urban environment.
The focus of this paper is given on a modified vehicle routing problem using electric vehicles instead. Given a
homogeneous fleet of depot-returning electric capacitated vehicles and a set of charge stations where vehicles
recharge their battery, the goal is to design a set of routes that satisfy the delivery requirements of a set of
geographically scattered customers. Each customer has a known demand for delivery and it must be serviced within
a predefined time window that denotes the earliest and the latest times to begin the service. Vehicles are charged at
the depot and can travel for a predefined distance before they need to visit a charging station in order to be fully
recharged. Finally, each customer must be visited only once by exactly one vehicle and the load of the vehicle must
not exceed its maximum capacity. The primary objective is to minimize the fleet size, while the secondary objective
is to minimize the distance traveled since the goal is to minimize the total routing cost.
Considering the Electric Vehicle Routing Problem (EVRP), the practical perspective embodies multiple
restrictions as well as operational constraints that enforced in practice for urban freight distribution using electric
454 Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459

vehicles. The EVRP is NP-hard since it is derived from the well-known Capacitated Vehicle Routing Problem (Toth
and Vigo, 2002) and thus, substantial computational effort is required for gaining high quality solutions even for
medium scale instances. Fig. 1 depicts a simple example of the EVRP where there is only one electric vehicle, 2
charge stations and a set of 9 customers. The aim of this paper is to develop a comprehensive model that captures the
realistic restrictions of the problem and also, to discuss and identify recent trends in transportation sector with
respect to the electric vehicles. To this end, essential insights for future development are presented in an attempt to
constitute the basis for future research inspired by industry needs.
The remainder of the paper is organized as follows: Section 2 provides an overview of the corresponding
literature and then, Section 3 presents the definition and the notation of the problem that describes the aspects of the
problem. Section 4 discusses and analyzes the problem variants that utilize electric vehicles as well as the proposed
solution approaches presented in literature providing also pointers for future research. Finally, in Section 5
conclusions are drawn giving a summary of the EVRP potentials.

Fig. 1. A route of the EVRP

2. Literature Review

The EVRP is introduced the latest years due to the commercialization of electric vehicles and their use in urban
freight distribution is imperative since the environmental protection and the minimization of the energy consumption
constitute major concerns for governments. Literature provides some indications regarding distances travelled for
urban deliveries and how this fits with the range of EVs. In the U.S. an average distance of 36 miles per day was
found by FHWA (2009, 2010), while a report by FedeX (Barnitt, 2011) concluded to a distance of 41,4 miles, as the
average travelled distance by delivery trucks. Both distances are significantly lower than the range achieved and
claimed by EV manufacturers. EV’s in CEP-services travel less than 40 km per day (Schönewolf, 2011). According
to University of Duisburg-Essen (2011), the following commercial sectors might incorporate EVs in their fleets:
CEP-Services, forwarding companies, and pharmaceutics, social services, newspaper and flower delivery.
Simulation and analysis of freight transport activities indicated cargo delivery with EVs in CEP, postal and public
cleaning services (University of Duisburg-Essen, 2011). Textile logistics also offer a large potential allowing multi-
shift delivery (Schönewolf, 2011). Moreover, CEP service providers often work with sub-contractors who own their
vehicle and use it privately.
Although the penetration of electric vehicles follows an accelerated pace, the literature about effective solution
approaches for electric vehicle routing and scheduling appeared limited and there is still much room for
contribution. Below, an overview of the studies encountered in literature is provided.
Artmeier et al. (2010) study the most economical plan of routing in terms of energy consumption rather than the
shortest one in terms of traveled distance and propose a generic solution framework for the Shortest Path Problem
considering trees of paths and an energy graph presenting the energy consumption. The proposed framework
encompasses four different strategies (Dijkstra, Expand, FIFO and Expand-distance) that determine the next vertex
Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459 455

to expand in an effort to obtain the energy-optimal route. It is the first attempt to introduce electric vehicles and the
main interest lies on the energy consumption considering acceleration and deceleration costs as well as hard and soft
constraints. The presented solution approach is implemented in a prototype navigation system that developed to
generate the optimum route in terms of energy consumption.
Later, Conrad and Figliozzi (2011) present a Rechargeable Vehicle Routing Problem with time windows where
vehicles can be recharged at customer locations during the service process and visits to customers are allowed only
within a predefined time window. This problem constitutes the basis for the studied EVRPTW and the main
objective is twofold; first, the minimization of the total number of vehicles and then, the minimization of the total
routing cost including traveled distance, service time and recharging time. Bounds on the average traveled distance
are established and based on the latter a regression analysis for the estimation of the average distance traveled is
occurred.
In the context of the renewable sources of energy, Erdoğan and Miller-Hooks (2012) describe a generic Green
Vehicle Routing Problem (GVRP) that encompasses the challenges associated with alternative fuel vehicles as the
electric vehicles are with the aim to minimize the total distance travelled by the vehicles. They developed a modified
Clarke and Wright construction heuristic based on geographical criteria (i.e. distance among customers) and a
density-based clustering algorithm based on spatial properties (i.e. distribution of customers in the space) followed
by a customized improvement technique that follows an inter- and intra- route exchange neighborhood search in
order to deal with the GVRP.
More recently, Schneider et al. (2014) study a vehicle routing problem with intermediate stops which considers
necessary visits at intermediate locations as the EVRP also requires. The main goal of the problem is to minimize
the sum of the total travel cost and the fixed vehicle cost and an efficient solution scheme introduced incorporating a
modified Clarke and Wright (1964) heuristic in order to generate the initial routes and also, adopts a metaheuristic
Adaptive Variable Neighborhood Search algorithm for further improvement. The latter employs both an exchange
and a relocate neighborhood structure, while the selection of customers is favored according to their previous
performance within the search process.
On the other hand, Barco et al. (2012) study the EVRP from the public transport perspective and present a case
study about carrying passengers from an airport to a hotel using electric vehicles and following the notation of the
Dial-a-Ride Problem (DARP). A scheme that coordinates the routing, charge scheduling and operating costs is
proposed taking into account the battery degradation as well as the recharging cost, while patterns allowing the
increase in the battery lifetime were also obtained.
Finally, Baouche at al. (2013) present an alternative study for the EVRP and efficient tools are developed to
minimize the energy consumption based on dynamic information in an effort to promote the use of electric vehicles
in practice. Although there are also many studies in literature including Suzuki (2011), Xiao et al. (2012) and Li
(2012) that are focused to energy consumption, their interest appears only on critical factors of the objective
function for energy consumption (i.e. travel speed, load and distance) without considering alternative sources of
energy (as electricity).

3. Problem Definition & Notation

According to the study of Davis and Figliozzi (2013), the most significant factors affecting commercial electric
vehicle competitiveness are the route feasibility, minimum fleet size, minimum traveled distance, charging level,
purchase costs and planning horizon. Inspired on the above and following the objective function of the classic
VRPTW, this research study presents a detailed model of the EVRP that considers realistic operational constraints
(i.e. vehicle capacity and time widows of customers) and captures the most important parameters for optimal routing
where electric vehicles exist. It is an attempt to provide a complete mathematical model for the EVRPTW that
incorporates the real-life processes and limitations that a company should contemplate for routing and scheduling a
fleet of electric vehicles.
The EVRPTW is defined as a complete and indirect graph G = (V, A), where V is a set of customers C = {c0, c1,
…, cn}, the charge stations S = {s1, …, sm} as well as the depot do that acts also as a charge station; and A is a set of
arcs A = {(i.j) ϵ V x V : i ≠ j}. Each arc is associated with a travel distance dij, and speed is assumed to be constant.
Given a set of vehicles K = {k1, k2, …, kk}, each customer i ϵ C is associated with a demand δi to be delivered, a
456 Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459

service time pi, and a time window [ei, li] that models the earliest ei and the latest li times during the day that service
can begin while for corresponding time window for the depot is [E, L]. During the service of the customers, all
vehicles must remain at the customer locations for a predefined amount of time (service time) and in case that a
vehicle arrives early; it has to wait wi time until its service. Similarly, vehicles remain parked at the stations for a
(charging) time hi during charging. A route starts from the depot, visits a number of customers and a number of
charging stations and returns to the depot as depicted in Fig 1. Note that each customer should be served by exactly
one vehicle that can visit him only once while the depot acts also as charge station and can be visited many times for
charging.
Given that vehicles visit the charging stations more than once, an extra set S’ = {sm+1, sm+2,…, sm+h} is used to
represent the multiple visits, and depot do is defined as a set D = {do+1, … do+t}. Hence, the new graph G’ = (V’, A’)
is created, where V’ = C‫׫‬S‫{׫‬do}‫׫‬S’‫ ׫‬D and A’ = {(i.j) ϵ V’ x V’ : i ≠ j}. Let xij be the binary variable that is defined
as 1 if the vehicle k travels from i to j and 0 otherwise; yi denotes the remaining electric energy after the visit at
customer i or at charge station s and depot do; qik denotes the time when service starts at customer i by vehicle k; τ
denotes the vehicle consumption rate, M the capacity and Q the battery level of charge; the mathematical
formulation is presented as follows:

min ¦ ¦ dij xijk (1)


kK ( i , j )Ac

Subject to:

¦¦x
kK jV c
ijk 1 i  C , k  K (2)

¦x
jV c
ijk  ¦ x jik
jV c
0 i V c, k  K
(3)

¦¦x
kK jV c
ijk d1 i V c \ C , k  K
(4)
xi , j ,k (qik  pi  tij  q jk ) d 0 k  K ,(i, j )  Ac
(5)
ei ( ¦ xijk ) d qik d li ( ¦ xijk ) k  K , i  C
jV c jV c
(6)

E d qik d L k  K , i {do } (7)

y j d yi  W dij xijk  Q(1  xijk ) i V c, j  C, k  K , i z j (8)

yi t min{W dio ,W (dij  d jo )} i  C, j  S Sc (9)

yi Q, i  do D S S'
(10)

¦G ¦ x
jC
j
iV c
ijk d M, k  K
(11)

xijk  0,1, i, j  Ac, k  K (12)

yi t 0, i  Ac (13)

qik t 0, i, j  Ac, k  K (14)


Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459 457

On the basis of the above, the objective function (1) indicates that there are two objectives, first to minimize the
number of vehicles required to service all customers and then, to minimize the total distance traveled. The (2)
constraint ensures that each customer is served by exactly one vehicle and constraint (3) present that the number of
the vehicles that arrive at a node is equal to the number of the vehicles that depart from it, and constraint (4) presents
the flow of a path followed by a vehicle k. Moreover, constraints (5), (6) and (7) represent the time windows
constraints and (11) the capacity constraint, where M are the units of loading capacity. However, based on Cordeau
et al. (2001), constraint (5) can be linearized as follows:

qik  pi  tij  q jk d (1  xijk ) Bij , i, j  Ac, k  K (15)

where Bij represents large constraints that could be replaced by max{li + pi + tij – ej, 0},  k‫ א‬K , (i,j)‫ א‬A’.
Constraint (8) represents the level of charge reduced according to the distance travelled; constraint (9) ensures that
there is enough remaining energy to return to the depot either directly or via a charge station; and constraint (10)
represents when a car is fully charged. Finally, the (12) is a binary constraint and either xijk = 1, in case that the
vehicle k serves the customer i and then the customer j, or xijk = 0 if not.
Regarding the energy consumption, the operating mode of the vehicle constitutes a crucial factor that should be
taken into consideration to estimate the yi remaining electric energy of each vehicle after the visit at a customer i.
Based on Davis and Figliozzi (2013), the energy consumption Ec of a vehicle with average speed υ (km/h) and
weight we (kg) due to the operating mode could be defined as:
U ˜ ul ˜ fv ˜X 3
Ec  f ˜X ˜ we ˜X  g ˜ we ˜X (16)
2
where ρ represents the air density (km/m3), ul is the coefficient of drag, g is the average road gradient (%) and fv is
the frontal area of the vehicle (m2) which depends on the average speed of the vehicle.

4. Current trends and motivations for future research

In this section the recent trends and observations about the EVRP are presented and discussed. Among these
special consideration is given on future research and directions are drawn with respect to industry needs and
technological developments.
To begin with, a shift towards more energy-efficient problems can be observed and an intense interest of the
research community belongs to EVRP. However, comparisons are not feasible since different objective functions
and constraints are encountered. This calls for an integrated study of the problem by researchers that requires a
systematic analysis of the real-world environment to capture the industry perspectives and exploit the scientific
background.
Towards this direction, dynamic information about travel times and customer requests could go the research a
step further in an effort to capture the dynamic nature of real problems encountered in practice. Studies dealing with
realistic constraints are still limited and the study of a real-world problem appeared in industry could enhance future
research. Another critical research pointer is the study of the cooling/heating usage for the freight EVs with respect
to the energy consumption. The energy effect of the cooling/heating usage could be examined under different
weather conditions (standard, cold and hot) since recent outcomes for passenger electric cars (Bütler and Winkler,
2013) prove the importance of this factor and particularly, it is observed that the heating usage appears a higher
impact on energy consumption compared to the cooling.
Similarly, the benchmark data sets used for evaluation are derived from instances previously generated for the
known capacitated VRP (such that instances generated by Christofides and Eilon (1969) and used for the EVRP by
Schneider et al. 2014) since industry information is not available. Therefore, a valuable pointer to future research is
the input of practitioners. Empirical studies dealing with small or large scale problems could play an important role
in the development of robust and effective solution methods.
Additionally, the proposed mathematical model could be enriched considering extra constraints such that a
heterogeneous fleet of vehicles, battery degradation as well as costs of recharge. Another possible modification of
458 Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459

studying the EVRP is to consider the possibility that customers own charging infrastructure offering fast charges for
a reduced transport fee of goods when delivery lasts longer than 20-30 minutes. This is an innovative idea due to the
restrictive use of the EVs and it could be really efficient for EV owners and cost-saving for customers.
Besides operational constraints, the practical perspective of a distribution – collection system is to route and
schedule these activities in a way that increases the possible synergies of combining pickup and delivery customers.
For this reason, merging products brought to the customers as well as products brought back to the depot is a
realistic case that seems to interest the logistics industry sector and it should be employed by researchers in near
future.
Due to the computational complexity of the EVRP, the current trend predominately belongs to approximate
methods (such that the Clarke and Wright heuristic, the density-based clustering algorithm and the Adaptive
Variable Neighborhood Search) that sacrifice the optimum solution for the sake of getting high quality solutions in a
significantly reduced amount of time. Comparison can be done by looking at the different results achieved for the
same benchmark instances. As it is already mentioned, comparisons for EVRP are not possible while in case of
GVRP, the modified benchmark data sets proposed by Erdoğan and Miller-Hooks (2012) are also used by Schneider
et al. (2014) and new best results are reported. Nonetheless, the evaluation of the proposed methods depends on
multiple criteria i.e. accuracy, speed, flexibility and simplicity (Cordeau et al. 2005). Thus, researchers should
consider a four-level analysis to develop more sophisticated solution procedures that embody approximation
algorithms, memory structures, operators and efficient search mechanisms of the solution space in a way that
enhances robustness and minimizes complexity.

5. Conclusions

Over the last years with increasing environmental consciousness, the use of electric vehicles has become critical
due to the economic and environmental importance for an effective and an energy-efficient urban freight
distribution. This paper draws essential observations about the recent trends of the EVRP and proposes a
mathematical model inspired of the known VRPTW. The main effort is to provide an outline of the literature and the
available solution frameworks identifying promising directions for future research that appear an intense interest in
practice. The latter reveal that moving towards more realistic and rich problems, real case studies are of great
interest and real-life data sets will provide the basis for the evaluation and the comparison of different solution
approaches. Finally, the development of more sophisticated solution schemes that render considerable efficiency
may also ensure future increase and effectiveness in use of the electric vehicles.

Acknowledgements

This research has been co-financed by the European Regional Development Fund (ERDF) in the context of the
project “SMILE – Smart green Innovative urban Logistics for Energy efficient Mediterranean cities” through the
program “STC Programme MED, Priority – Objective 2-2”.

References

Artmeier, A., Haselmayr, J., Leucker, M., Sachenbacher, M., 2010. The shortest path problem revisited: Optimal routing for electric vehicles.
In KI 2010: Advances in Artificial Intelligence, (Ed.) Springer Berlin Heidelberg, pp. 309-316.
Barco, J., Guerra, A., Muñoz, L., Quijano, N. 2013. Optimal Routing and Scheduling of Charge for Electric Vehicles: Case Study. arXiv preprint
arXiv:1310.0145.
Barnes, J. W., Wiley, V. D., Moore, J. T., Ryer, D. M., 2004. Solving the aerial fleet refueling problem using group theoretic tabu
search. Mathematical and computer modelling 39(6), 617-640.
Barnitt, R., 2011. FedEx Express Gasoline Hybrid Electric Delivery Truck Evaluation: 12 Month Report, Colorado.
Baouche, F., Trigui, R., El Faouzi, N. E., & Billot, R., 2013. Energy Consumption Assessment For Electric Vehicles. In International symposium
on recent advances in transport modeling.
Bütler, T., Hannes, W., 2013. Energy consumption of battery electric vehicles. EMPA - Research Institute, Laboratory for internal combustion
engines.
Christofides, N. , Eilon, S., 1969. An algorithm for thr vehicle-dispatching problem. Operational Research Quarterly 20(3), 309-318.
Clarke, G., Wright, J. W., 1964. Scheduling of vehicles from a central depot to a number of delivery points. Operations research 12(4), 568-581.
Anagnostopoulou Afroditi et al. / Transportation Research Procedia 3 (2014) 452 – 459 459

Conrad, R. G., Figliozzi, M. A., 2011. The recharging vehicle routing problem. In Proceedings Industrial Engineering Research Conference.
Reno, NV.
Cordeau, J. F., Desaulniers, G., Desrosiers, J., Solomon, M. M., Soumis, F., 2001. VRP with time windows. In Toth, P. and Vigo, D. The vehicle
routing problem, SIAM Monographs on Discrete Mathematics and Applications, Philadelphia, 157-193.
Cordeau, J. F., Gendreau, M., Hertz, A., Laporte, G., Sormany, J. S., 2005. New heuristics for the vehicle routing problem, (Ed.) Springer US, pp.
279-297.
Davis, B. A., Figliozzi, M. A., 2013. A methodology to evaluate the competitiveness of electric delivery trucks. Transportation Research Part E:
Logistics and Transportation Review 49(1), 8-23.
Ehrler, V., Hebes P., 2012. Electromobility for city logistics – The solution to urban transport collapse? An analysis beyond theory, Procedia –
Social and Behavioral Sciences 48, 786 – 795.
Erdoğan, S., Miller-Hooks, E., 2012. A green vehicle routing problem. Transportation Research Part E: Logistics and Transportation Review
48(1), 100-114.
FHWA, 2009. Highway statistics 2009, table VM-1.
Feng, W. and Figliozzi, M., 2013. An economic and technological analysis of the key factors affecting the competitiveness of electric commercial
vehicles: A case study from the USA market. Transportation Research Part C: Emerging Technologies, Volume 26, January 2013, 135–145.
Jüchter, A., 2011. Projektsteckbrief E-City Logistik - Teilvorhaben E-KEP. Berlin
Koupal, J. W., 2001. Air conditioning activity effects in MOBILE6. M6. ACE, 1.
Li, J., 2012. Vehicle Routing Problem with Time Windows for Reducing Fuel Consumption. Journal of Computers, 7(12), 3020-3027.
Schneider, M., Stenger, A., Hof, J., Professorship, D. S. E. A., Darmstadt, T. U., 2014. An Adaptive VNS Algorithm for Vehicle Routing
Problems with Intermediate Stops (No. 63500). Darmstadt Technical University, Department of Business Administration, Economics and
Law, Institute for Business Studies (BWL).
Schönewolf, W, 2011. E-City-Logistik, Status-Seminar Elektromobilität BerlinBrandenburg, Berlin, 25/10/2011.
Suzuki, Y., 2011. A new truck-routing approach for reducing fuel consumption and pollutants emission. Transportation Research Part D:
Transport and Environment, 16(1), 73-77.
Toth, P., Vigo, D., 2002. The vehicle routing problem (Vol. 9). Siam.
University of Duisburg-Essen., 2011. Schlussbericht der Universität Duisburg-Essen zum Forschungsvorhaben colognE-mobil - Simulation und
Begleitforschung Modellregion Rhein-Ruhr.
USEPA., 2010. MOVES 2010 Highway Vehicle Temperature, Humidity, Air Conditioning, and Inspection and Maintenance Adjustments. s.l. :
EPA-420-R-10-027.
Xiao, Y., Zhao, Q., Kaku, I., Xu, Y., 2012. Development of a fuel consumption optimization model for the capacitated vehicle routing
problem. Computers & Operations Research, 39(7), 1419-1431.